2021-08-30 distributed cluster

1、 Prepare four virtual machines

The user names are node1,node2,node3,node4

Enter the command
cd /etc/sysconfig/network-scripts/

modify vim vim ifcfg-eth0, In order to make IP Different settings are static, as shown in the following figure ：

2、 Realize the secret free login of four virtual machines

（1） First, execute on all four servers ：
ssh-keygen -t dsa -P ‘’ -f ~/.ssh/id_dsa
（2） stay node1 Admiral node1 Copy the public key of to authorized_keys in ：
id_dsa.pub ： Public key
known_hosts Private key
cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys
（3） Copy the file to node2：
scp ~/.ssh/authorized_keys node2:/root/.ssh/
（4） stay node2 Lieutenant general node2 The public key of is appended to authorized_keys in ：
cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys
（5） Copy the file to node3：
scp ~/.ssh/authorized_keys node3:/root/.ssh/
（6） stay node3 Lieutenant general node3 The public key of is appended to authorized_keys in ：
cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys
（7） Copy the file to node4：
scp ~/.ssh/authorized_keys node4:/root/.ssh/
（8） stay node4 Lieutenant general node4 The public key of is appended to authorized_keys in ：
cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys
（9） Copy the file to node1、node2、node3：
scp ~/.ssh/authorized_keys node1:/root/.ssh/
scp ~/.ssh/authorized_keys node2:/root/.ssh/
scp ~/.ssh/authorized_keys node3:/root/.ssh/
Be careful ： Use scp When it's time to vim /etc/hosts file , That is, the host that needs to log in without secret IP And the user name

3、JDK Install environment variable configuration

node1-node4

mkdir /opt/apps

take jdk-8u221-linux-x64.rpm Upload to node1/opt/apps
take /opt/apps Under the jdk.rpm scp To node2、node3、node4 In the corresponding directory of

scp jdk-8u221-linux-x64.rpm node2:/opt/apps
scp jdk-8u221-linux-x64.rpm node3:/opt/apps
scp jdk-8u221-linux-x64.rpm node4:/opt/apps

stay node1、node2、node3、node4 Installation on jdk And configuration profile file

rpm -ivh jdk-8u221-linux-x64.rpm

node1 Modify environment variables on

vim /etc/profile
export JAVA_HOME=/usr/java/default
export PATH=$PATH:$JAVA_HOME/bin
source /etc/profile

take node1 Of /etc/profile copy to node2、node3、node4 Go up and execute . /etc/profile

scp /etc/profile node[234]:`pwd`

4、ZooKeeper Cluster building

take ZooKeeper.tar.gz Upload to node2
b) Unzip to /opt

tar -zxvf ZooKeeper-3.4.6.tar.gz -C /opt

c) Configure environment variables ：

export ZOOKEEPER_HOME=/opt/zookeeper-3.4.6
export PATH=$PATH:$JAVA_HOME/bin:$ZOOKEEPER_HOME/bin

then ./etc/profile Make the configuration work
Finally, the file scp To node3 and node4 On , And separately ./etc/profile Make the configuration work .

scp /etc/profile node3:/etc/
scp /etc/profile node4:/etc/

d) To $ZooKeeper_HOME/conf Next
Copy zoo_sample.cfg by zoo.cfg

cp zoo_sample.cfg zoo.cfg

e) edit zoo.cfg
Add ： Parameter description

tickTime=2000 # The interval between heartbeats , Company ： millisecond 
dataDir=/opt/zookeeper-3.4.6/data #ZooKeeper  The directory where the data is stored 
dataLogDir=/var/bjsxt/zookeeper/datalog # Log directory 
clientPort=2181
initLimit=5
syncLimit=2
server.1=server2:2881:3881
server.2=server3:2881:3881
server.3=node4:2881:3881 #observer（ Indicates that the corresponding node does not participate in voting ）

clientPort： Client connection ZooKeeper The port of the server ,ZooKeeper Will listen to this port , Pick up
Access request from client .
initLimit： This configuration item is used for configuration ZooKeeper Accept clients （ The client mentioned here is not a user
Connect ZooKeeper The client side of the server , It is ZooKeeper Connect to the server cluster Leader Of
Follower The server ） The maximum number of heartbeat intervals that can be tolerated when initiating a connection . When it has exceeded 5 A heart
Jump time （ That is to say tickTime） After the length ZooKeeper The server has not received a return message from the client ,
So it indicates that the client connection failed . The total length of time is zero 52000=10 second
syncLimit： This configuration item identity Leader And Follower Sending messages between , Length of request and reply time ,
No more than that tickTime Length of time , The total length of time is zero 22000=4 second
server.A=B：C：D： Its in A It's a number , Which server is this ;B It's this server
Of ip Address ;C It represents the server and the cluster Leader The port on which the server exchanges information ;D It means
In case of Leader The server is down , You need a port to re-election , Pick a new one
Leader, And this port is the port that the servers communicate with each other during the election . If it is the configurator of the pseudo cluster
type , because B Is the same , So different ZooKeeper The instance communication port number cannot be the same , So give them points
With different port numbers .
Add as follows ：

server.1=node2:2881:3881
server.2=node3:2881:3881
server.3=node4:2881:3881

modify

dataDir=/opt/ZooKeeper-3.4.6/data

f) establish /opt/ZooKeeper-3.4.6/data Catalog , And put a file in the directory ：myid
stay myid Write down the current ZooKeeper The number of

mkdir /opt/zookeeper-3.4.6/data
mkdir /var/bjsxt/zookeeper/datalog
echo 1 > /opt/zookeeper-3.4.6/data/myid

g) Configure the ZooKeeper copy to node3、node4 On

scp -r ZooKeeper-3.4.6/ node3:/opt/
scp -r ZooKeeper-3.4.6/ node4:/opt/

h) stay node3 and node4 I'm going to change ：myid
node3

echo 2 > /opt/ZooKeeper-3.4.6/data/myid

node4

echo 3 > /opt/ZooKeeper-3.4.6/data/myid

i) To start, respectively, ZooKeeper

zkServer.sh start  start-up  zk
zkServer.sh stop  stop it  zk
zkServer.sh status  see  zk  state 
zkServer.sh start|stop|status

j) close ZooKeeper

zkServer.sh stop

l) Connect ZooKeeper
zkCli.sh node2、node3、node4 Fine
m) sign out zkCli.sh command
quit

Provide cluster mode services
Atomicity
Accurate feedback on success or failure
Uniformity
Every server There is a unified data view
Usability
Node failure does not affect the use Less than half the downtime , That's all right.
Network partition / Split brain ： More than half pass
Who is older myid：
1 2 3 （ At the same time to start ）
1 2 3 （ Start one by one ）
Who has new data ：
3 3 4 Business id（zxid） Big ones are leaders
3 Taiwan machine Hang one 2>3/2
5 Taiwan machine Hang up 1 platform 4>5/2（ Can work ） hang 3 platform 2！>5/2（ Clusters don't work ）
Sequence ：FIFO
The master-slave model
Leader Follower+
Read more than a writing
Leader Responsible for adding, deleting and modifying ,Follower Read and vote
State of the cluster
Election model Anqinei
Broadcast mode Outside
Server state
LOOKING： At present Server I do not know! leader Who is it? , Searching for
LEADING： At present Server It's an elected leader
FOLLOWING：leader It has been elected , At present Server Keep up with it
Division of labor between master and slave
The leader （leader）
Responsible for the initiation and resolution of the vote , Update system status ,（ Additions and deletions ）
Learners' （learner）
Including followers （follower） And the observer （observer）,follower Used to accept clients
Request and return results to the client , Participate in voting during operation
Observer
Can accept client connections , Forward the write request to leader, but observer Not in the voting process ,
Sync only leader The state of ,observer The goal is to extend the system , Improve read speed
client （client）
Request originator